Anna Vorontsova

Danila Rukhovich, Anna Vorontsova, Anton Konushin

Recently, sparse 3D convolutions have changed 3D object detection. Performing on par with the voting-based approaches, 3D CNNs are memory-efficient and scale to large scenes better. However, there is still room for improvement. With a conscious, practice-oriented approach to problem-solving, we analyze the performance of such methods and localize the weaknesses. Applying modifications that resolve the found issues one by one, we end up with TR3D: a fast fully-convolutional 3D object detection model trained end-to-end, that achieves state-of-the-art results on the standard benchmarks, ScanNet v2, SUN RGB-D, and S3DIS. Moreover, to take advantage of both point cloud and RGB inputs, we introduce an early fusion of 2D and 3D features. We employ our fusion module to make conventional 3D object detection methods multimodal and demonstrate an impressive boost in performance. Our model with early feature fusion, which we refer to as TR3D+FF, outperforms existing 3D object detection approaches on the SUN RGB-D dataset. Overall, besides being accurate, both TR3D and TR3D+FF models are lightweight, memory-efficient, and fast, thereby marking another milestone on the way toward real-time 3D object detection. Code is available at https://github.com/SamsungLabs/tr3d .

Maksim Kolodiazhnyi, Anna Vorontsova, Anton Konushin et al.

Most 3D instance segmentation methods exploit a bottom-up strategy, typically including resource-exhaustive post-processing. For point grouping, bottom-up methods rely on prior assumptions about the objects in the form of hyperparameters, which are domain-specific and need to be carefully tuned. On the contrary, we address 3D instance segmentation with a TD3D: the pioneering cluster-free, fully-convolutional and entirely data-driven approach trained in an end-to-end manner. This is the first top-down method outperforming bottom-up approaches in 3D domain. With its straightforward pipeline, it demonstrates outstanding accuracy and generalization ability on the standard indoor benchmarks: ScanNet v2, its extension ScanNet200, and S3DIS, as well as on the aerial STPLS3D dataset. Besides, our method is much faster on inference than the current state-of-the-art grouping-based approaches: our flagship modification is 1.9x faster than the most accurate bottom-up method, while being more accurate, and our faster modification shows state-of-the-art accuracy running at 2.6x speed. Code is available at https://github.com/SamsungLabs/td3d .

Maksim Kolodiazhnyi, Anna Vorontsova, Matvey Skripkin et al.

Growing customer demand for smart solutions in robotics and augmented reality has attracted considerable attention to 3D object detection from point clouds. Yet, existing indoor datasets taken individually are too small and insufficiently diverse to train a powerful and general 3D object detection model. In the meantime, more general approaches utilizing foundation models are still inferior in quality to those based on supervised training for a specific task. In this work, we propose \ours{}, a simple yet effective 3D object detection model, which is trained on a mixture of indoor datasets and is capable of working in various indoor environments. By unifying different label spaces, \ours{} enables learning a strong representation across multiple datasets through a supervised joint training scheme. The proposed network architecture is built upon a vanilla transformer encoder, making it easy to run, customize and extend the prediction pipeline for practical use. Extensive experiments demonstrate that \ours{} obtains significant gains over existing 3D object detection methods in 6 indoor benchmarks: ScanNet (+1.1 mAP50), ARKitScenes (+19.4 mAP25), S3DIS (+9.1 mAP50), MultiScan (+9.3 mAP50), 3RScan (+3.2 mAP50), and ScanNet++ (+2.7 mAP50). Code is available at https://github.com/filapro/unidet3d .

Maxim Kolodiazhnyi, Anna Vorontsova, Anton Konushin et al.

Semantic, instance, and panoptic segmentation of 3D point clouds have been addressed using task-specific models of distinct design. Thereby, the similarity of all segmentation tasks and the implicit relationship between them have not been utilized effectively. This paper presents a unified, simple, and effective model addressing all these tasks jointly. The model, named OneFormer3D, performs instance and semantic segmentation consistently, using a group of learnable kernels, where each kernel is responsible for generating a mask for either an instance or a semantic category. These kernels are trained with a transformer-based decoder with unified instance and semantic queries passed as an input. Such a design enables training a model end-to-end in a single run, so that it achieves top performance on all three segmentation tasks simultaneously. Specifically, our OneFormer3D ranks 1st and sets a new state-of-the-art (+2.1 mAP50) in the ScanNet test leaderboard. We also demonstrate the state-of-the-art results in semantic, instance, and panoptic segmentation of ScanNet (+21 PQ), ScanNet200 (+3.8 mAP50), and S3DIS (+0.8 mIoU) datasets.

Nikolay Patakin, Mikhail Romanov, Anna Vorontsova et al.

Nowadays, robotics, AR, and 3D modeling applications attract considerable attention to single-view depth estimation (SVDE) as it allows estimating scene geometry from a single RGB image. Recent works have demonstrated that the accuracy of an SVDE method hugely depends on the diversity and volume of the training data. However, RGB-D datasets obtained via depth capturing or 3D reconstruction are typically small, synthetic datasets are not photorealistic enough, and all these datasets lack diversity. The large-scale and diverse data can be sourced from stereo images or stereo videos from the web. Typically being uncalibrated, stereo data provides disparities up to unknown shift (geometrically incomplete data), so stereo-trained SVDE methods cannot recover 3D geometry. It was recently shown that the distorted point clouds obtained with a stereo-trained SVDE method can be corrected with additional point cloud modules (PCM) separately trained on the geometrically complete data. On the contrary, we propose GP$^{2}$, General-Purpose and Geometry-Preserving training scheme, and show that conventional SVDE models can learn correct shifts themselves without any post-processing, benefiting from using stereo data even in the geometry-preserving setting. Through experiments on different dataset mixtures, we prove that GP$^{2}$-trained models outperform methods relying on PCM in both accuracy and speed, and report the state-of-the-art results in the general-purpose geometry-preserving SVDE. Moreover, we show that SVDE models can learn to predict geometrically correct depth even when geometrically complete data comprises the minor part of the training set.

Danil Galeev, Polina Popenova, Anna Vorontsova et al.

Recent advances in interactive segmentation (IS) allow speeding up and simplifying image editing and labeling greatly. The majority of modern IS approaches accept user input in the form of clicks. However, using clicks may require too many user interactions, especially when selecting small objects, minor parts of an object, or a group of objects of the same type. In this paper, we consider such a natural form of user interaction as a loose contour, and introduce a contour-based IS method. We evaluate the proposed method on the standard segmentation benchmarks, our novel UserContours dataset, and its subset UserContours-G containing difficult segmentation cases. Through experiments, we demonstrate that a single contour provides the same accuracy as multiple clicks, thus reducing the required amount of user interactions.

Anna Sokolova, Filipp Nikitin, Anna Vorontsova et al.

Processing large indoor scenes is a challenging task, as scan registration and camera trajectory estimation methods accumulate errors across time. As a result, the quality of reconstructed scans is insufficient for some applications, such as visual-based localization and navigation, where the correct position of walls is crucial. For many indoor scenes, there exists an image of a technical floorplan that contains information about the geometry and main structural elements of the scene, such as walls, partitions, and doors. We argue that such a floorplan is a useful source of spatial information, which can guide a 3D model optimization. The standard RGB-D 3D reconstruction pipeline consists of a tracking module applied to an RGB-D sequence and a bundle adjustment (BA) module that takes the posed RGB-D sequence and corrects the camera poses to improve consistency. We propose a novel optimization algorithm expanding conventional BA that leverages the prior knowledge about the scene structure in the form of a floorplan. Our experiments on the Redwood dataset and our self-captured data demonstrate that utilizing floorplan improves accuracy of 3D reconstructions.

Anton Konushin, Nikita Drozdov, Bulat Gabdullin et al.

Layout estimation and 3D object detection are two fundamental tasks in indoor scene understanding. When combined, they enable the creation of a compact yet semantically rich spatial representation of a scene. Existing approaches typically rely on point cloud input, which poses a major limitation since most consumer cameras lack depth sensors and visual-only data remains far more common. We address this issue with TUN3D, the first method that tackles joint layout estimation and 3D object detection in real scans, given multi-view images as input, and does not require ground-truth camera poses or depth supervision. Our approach builds on a lightweight sparse-convolutional backbone and employs two dedicated heads: one for 3D object detection and one for layout estimation, leveraging a novel and effective parametric wall representation. Extensive experiments show that TUN3D achieves state-of-the-art performance across three challenging scene understanding benchmarks: (i) using ground-truth point clouds, (ii) using posed images, and (iii) using unposed images. While performing on par with specialized 3D object detection methods, TUN3D significantly advances layout estimation, setting a new benchmark in holistic indoor scene understanding. Code is available at https://github.com/col14m/tun3d .

Danila Rukhovich, Anna Vorontsova, Anton Konushin

Recently, promising applications in robotics and augmented reality have attracted considerable attention to 3D object detection from point clouds. In this paper, we present FCAF3D - a first-in-class fully convolutional anchor-free indoor 3D object detection method. It is a simple yet effective method that uses a voxel representation of a point cloud and processes voxels with sparse convolutions. FCAF3D can handle large-scale scenes with minimal runtime through a single fully convolutional feed-forward pass. Existing 3D object detection methods make prior assumptions on the geometry of objects, and we argue that it limits their generalization ability. To get rid of any prior assumptions, we propose a novel parametrization of oriented bounding boxes that allows obtaining better results in a purely data-driven way. The proposed method achieves state-of-the-art 3D object detection results in terms of mAP@0.5 on ScanNet V2 (+4.5), SUN RGB-D (+3.5), and S3DIS (+20.5) datasets. The code and models are available at https://github.com/samsunglabs/fcaf3d.

Danila Rukhovich, Anna Vorontsova, Anton Konushin

In this paper, we introduce the task of multi-view RGB-based 3D object detection as an end-to-end optimization problem. To address this problem, we propose ImVoxelNet, a novel fully convolutional method of 3D object detection based on monocular or multi-view RGB images. The number of monocular images in each multi-view input can variate during training and inference; actually, this number might be unique for each multi-view input. ImVoxelNet successfully handles both indoor and outdoor scenes, which makes it general-purpose. Specifically, it achieves state-of-the-art results in car detection on KITTI (monocular) and nuScenes (multi-view) benchmarks among all methods that accept RGB images. Moreover, it surpasses existing RGB-based 3D object detection methods on the SUN RGB-D dataset. On ScanNet, ImVoxelNet sets a new benchmark for multi-view 3D object detection. The source code and the trained models are available at https://github.com/saic-vul/imvoxelnet.

Andrey Moskalenko, Vlad Shakhuro, Anna Vorontsova et al.

Interactive segmentation methods rely on user inputs to iteratively update the selection mask. A click specifying the object of interest is arguably the most simple and intuitive interaction type, and thereby the most common choice for interactive segmentation. However, user clicking patterns in the interactive segmentation context remain unexplored. Accordingly, interactive segmentation evaluation strategies rely more on intuition and common sense rather than empirical studies (e.g., assuming that users tend to click in the center of the area with the largest error). In this work, we conduct a real user study to investigate real user clicking patterns. This study reveals that the intuitive assumption made in the common evaluation strategy may not hold. As a result, interactive segmentation models may show high scores in the standard benchmarks, but it does not imply that they would perform well in a real world scenario. To assess the applicability of interactive segmentation methods, we propose a novel evaluation strategy providing a more comprehensive analysis of a model's performance. To this end, we propose a methodology for finding extreme user inputs by a direct optimization in a white-box adversarial attack on the interactive segmentation model. Based on the performance with such adversarial user inputs, we assess the robustness of interactive segmentation models w.r.t click positions. Besides, we introduce a novel benchmark for measuring the robustness of interactive segmentation, and report the results of an extensive evaluation of dozens of models.

Maksim Kolodiazhnyi, Denis Tarasov, Dmitrii Zhemchuzhnikov et al.

Computer-Aided Design (CAD) plays a central role in engineering and manufacturing, making it possible to create precise and editable 3D models. Using a variety of sensor or user-provided data as inputs for CAD reconstruction can democratize access to design applications. However, existing methods typically focus on a single input modality, such as point clouds, images, or text, which limits their generalizability and robustness. Leveraging recent advances in vision-language models (VLM), we propose a multi-modal CAD reconstruction model that simultaneously processes all three input modalities. Inspired by large language model (LLM) training paradigms, we adopt a two-stage pipeline: supervised fine-tuning (SFT) on large-scale procedurally generated data, followed by reinforcement learning (RL) fine-tuning using online feedback, obtained programatically. Furthermore, we are the first to explore RL fine-tuning of LLMs for CAD tasks demonstrating that online RL algorithms such as Group Relative Preference Optimization (GRPO) outperform offline alternatives. In the DeepCAD benchmark, our SFT model outperforms existing single-modal approaches in all three input modalities simultaneously. More importantly, after RL fine-tuning, cadrille sets new state-of-the-art on three challenging datasets, including a real-world one.

Polina Karpikova, Andrei Spiridonov, Anna Vorontsova et al.

Self-portraits captured from a short distance might look unnatural or even unattractive due to heavy distortions making facial features malformed, and ill-placed head poses. In this paper, we propose SUPER, a novel method of eliminating distortions and adjusting head pose in a close-up face crop. We perform 3D GAN inversion for a facial image by optimizing camera parameters and face latent code, which gives a generated image. Besides, we estimate depth from the obtained latent code, create a depth-induced 3D mesh, and render it with updated camera parameters to obtain a warped portrait. Finally, we apply the visibility-based blending so that visible regions are reprojected, and occluded parts are restored with a generative model. Experiments on face undistortion benchmarks and on our self-collected Head Rotation dataset (HeRo), show that SUPER outperforms previous approaches both qualitatively and quantitatively, opening new possibilities for photorealistic selfie editing.

Anna Sokolova, Anna Vorontsova, Bulat Gabdullin et al.

Leveraging 3D semantics for direct 3D reconstruction has a great potential yet unleashed. For instance, by assuming that walls are vertical, and a floor is planar and horizontal, we can correct distorted room shapes and eliminate local artifacts such as holes, pits, and hills. In this paper, we propose FAWN, a modification of truncated signed distance function (TSDF) reconstruction methods, which considers scene structure by detecting walls and floor in a scene, and penalizing the corresponding surface normals for deviating from the horizontal and vertical directions. Implemented as a 3D sparse convolutional module, FAWN can be incorporated into any trainable pipeline that predicts TSDF. Since FAWN requires 3D semantics only for training, no additional limitations on further use are imposed. We demonstrate, that FAWN-modified methods use semantics more effectively, than existing semantic-based approaches. Besides, we apply our modification to state-of-the-art TSDF reconstruction methods, and demonstrate a quality gain in SCANNET, ICL-NUIM, TUM RGB-D, and 7SCENES benchmarks.

Mikhail Artemyev, Anna Vorontsova, Anna Sokolova et al.

The majority of modern single-view depth estimation methods predict relative depth and thus cannot be directly applied in many real-world scenarios, despite impressive performance in the benchmarks. Moreover, single-view approaches cannot guarantee consistency across a sequence of frames. Consistency is typically addressed with test-time optimization of discrepancy across views; however, it takes hours to process a single scene. In this paper, we present MEDeA, an efficient multi-view test-time depth adjustment method, that is an order of magnitude faster than existing test-time approaches. Given RGB frames with camera parameters, MEDeA predicts initial depth maps, adjusts them by optimizing local scaling coefficients, and outputs temporally-consistent depth maps. Contrary to test-time methods requiring normals, optical flow, or semantics estimation, MEDeA produces high-quality predictions with a depth estimation network solely. Our method sets a new state-of-the-art on TUM RGB-D, 7Scenes, and ScanNet benchmarks and successfully handles smartphone-captured data from ARKitScenes dataset.

Mikhail Romanov, Nikolay Patatkin, Anna Vorontsova et al.

Single-view depth estimation (SVDE) plays a crucial role in scene understanding for AR applications, 3D modeling, and robotics, providing the geometry of a scene based on a single image. Recent works have shown that a successful solution strongly relies on the diversity and volume of training data. This data can be sourced from stereo movies and photos. However, they do not provide geometrically complete depth maps (as disparities contain unknown shift value). Therefore, existing models trained on this data are not able to recover correct 3D representations. Our work shows that a model trained on this data along with conventional datasets can gain accuracy while predicting correct scene geometry. Surprisingly, only a small portion of geometrically correct depth maps are required to train a model that performs equally to a model trained on the full geometrically correct dataset. After that, we train computationally efficient models on a mixture of datasets using the proposed method. Through quantitative comparison on completely unseen datasets and qualitative comparison of 3D point clouds, we show that our model defines the new state of the art in general-purpose SVDE.

Igor Slinko, Anna Vorontsova, Dmitry Zhukov et al.

Learning-based visual odometry and SLAM methods demonstrate a steady improvement over past years. However, collecting ground truth poses to train these methods is difficult and expensive. This could be resolved by training in an unsupervised mode, but there is still a large gap between performance of unsupervised and supervised methods. In this work, we focus on generating synthetic data for deep learning-based visual odometry and SLAM methods that take optical flow as an input. We produce training data in a form of optical flow that corresponds to arbitrary camera movement between a real frame and a virtual frame. For synthesizing data we use depth maps either produced by a depth sensor or estimated from stereo pair. We train visual odometry model on synthetic data and do not use ground truth poses hence this model can be considered unsupervised. Also it can be classified as monocular as we do not use depth maps on inference. We also propose a simple way to convert any visual odometry model into a SLAM method based on frame matching and graph optimization. We demonstrate that both the synthetically-trained visual odometry model and the proposed SLAM method build upon this model yields state-of-the-art results among unsupervised methods on KITTI dataset and shows promising results on a challenging EuRoC dataset.

David Prokhorov, Dmitry Zhukov, Olga Barinova et al.

Simultaneous localization and mapping (SLAM) is an essential component of robotic systems. In this work we perform a feasibility study of RGB-D SLAM for the task of indoor robot navigation. Recent visual SLAM methods, e.g. ORBSLAM2 \cite{mur2017orb}, demonstrate really impressive accuracy, but the experiments in the papers are usually conducted on just a few sequences, that makes it difficult to reason about the robustness of the methods. Another problem is that all available RGB-D datasets contain the trajectories with very complex camera motions. In this work we extensively evaluate ORBSLAM2 to better understand the state-of-the-art. First, we conduct experiments on the popular publicly available datasets for RGB-D SLAM across the conventional metrics. We perform statistical analysis of the results and find correlations between the metrics and the attributes of the trajectories. Then, we introduce a new large and diverse HomeRobot dataset where we model the motions of a simple home robot. Our dataset is created using physically-based rendering with realistic lighting and contains the scenes composed by human designers. It includes thousands of sequences, that is two orders of magnitude greater than in previous works. We find that while in many cases the accuracy of SLAM is very good, the robustness is still an issue.

Pavel Kirsanov, Airat Gaskarov, Filipp Konokhov et al.

We present a novel dataset for training and benchmarking semantic SLAM methods. The dataset consists of 200 long sequences, each one containing 3000-5000 data frames. We generate the sequences using realistic home layouts. For that we sample trajectories that simulate motions of a simple home robot, and then render the frames along the trajectories. Each data frame contains a) RGB images generated using physically-based rendering, b) simulated depth measurements, c) simulated IMU readings and d) ground truth occupancy grid of a house. Our dataset serves a wider range of purposes compared to existing datasets and is the first large-scale benchmark focused on the mapping component of SLAM. The dataset is split into train/validation/test parts sampled from different sets of virtual houses. We present benchmarking results forboth classical geometry-based and recent learning-based SLAM algorithms, a baseline mapping method, semantic segmentation and panoptic segmentation.

Igor Slinko, Anna Vorontsova, Filipp Konokhov et al.

Optical Flow (OF) and depth are commonly used for visual odometry since they provide sufficient information about camera ego-motion in a rigid scene. We reformulate the problem of ego-motion estimation as a problem of motion estimation of a 3D-scene with respect to a static camera. The entire scene motion can be represented as a combination of motions of its visible points. Using OF and depth we estimate a motion of each point in terms of 6DoF and represent results in the form of motion maps, each one addressing single degree of freedom. In this work we provide motion maps as inputs to a deep neural network that predicts 6DoF of scene motion. Through our evaluation on outdoor and indoor datasets we show that utilizing motion maps leads to accuracy improvement in comparison with naive stacking of depth and OF. Another contribution of our work is a novel network architecture that efficiently exploits motion maps and outperforms learnable RGB/RGB-D baselines.